206 PROCEEDINGS OF THE AMERICAN ACADEMY. 



in use in Ionic* It is pretty clear that the action of eireiprjTcu is here 

 aoristic. 



A very remarkable use of the present infinitive after eW, if it really is 

 after «W, has been cited above, p. 179. Assuming that eas is bere the 

 temporal conjunction, we should certainly expect the aorist. As was 

 mentioned above, Gildersleeve calls attention to the plural nXoloicn. But 



cf. 2, 179 : to (fiopTia efiff nepidytiv ev (Bupiai nepl to Ae'Aru, pe'xpi °^ " 7r ' 

 koito (s NavKpariv ; anil 2, 102: tov eXtyov ol iptts npooTov pei> 7rXoioiai 

 pciKpola i opprjOtVTa . . . tovs . . . KaTOLKrjpevovs Karao-rpe(p(a8aL, is b irXtovrd 

 piv Trpoaio diriKeo-dai is 6dXao-o-av, kt.X. In this connection it is not with- 

 out significance that, in 67 cases of dinkvUa8ai in clauses of antecedence, 

 the aorist is used in every case. As a matter of fact the imperfect of this 

 verb is very rare in Herodotus. Schweighauser does not cite it at all. t" 

 The difficulty, if not the impossibility, of explaining the present, is a 

 strong indication that ews is here not temporal at all. 



The use of vip(crdat in 7, 171, after tare: Xipov re koi Xotpov ytvecrdat 

 . . . , ecrre to bevTepov eprjpaideiarrjs KprjTrjs peTa tSov vjroXolncov Tpirovs aliTrjv 



vlv vepeadai KprjTas, has been explained by Gildersleeve t and Fuchs.§ 

 epr]pui6fio-r]s forms the virtual conclusion of ecrre, and the vvv . . . vi\ieo-Qai 

 looks very much like a sudden and unannounced shift from a subordi- 

 nate to a main construction. 



The imperfect is found in clauses with is 5 eight times || (1, 158, 196 ; 



3, 48 ; 4, 160, 1 60 2 ; 6, 1 13 ; 9, 55, 94). In 3, 48 (o'ixovto dTroXiTrovres), and 



4, 160 (dTroXnrovTfs o'ixovto), an aorist participle occurs, and o'lxovto is itself 



aoristic. 1, 158 : 'ApiaTob'iKos . . . ecr^e prj noirjaai ravTa Kvpaiovs, . . . (g 

 o to deirepov . . . ineiprjaopevoi rjio~av uXXol deonponoi, k.t.X. Note that tipt has 



no aorist ; furthermore, this passage seems to give evidence, in addition 

 to the conclusions reached above (p. 191), that the preterite of dpi is in 

 Herodotus sometimes, at least, aoristic. 1, 196: kci\ Tavrrjv dveKr]pvo-o-e 

 . . . is o rw to iXdxio~Tov imio-Tapevu) npoatKeiTo. The verb here used has 

 no aorist. In 6, 113, an aorist participle is used with the imperfect. 



In 9, 94: \_Evr]viu>~\ irapi^ovTo Kal Xoyovs iiXXovs inoievvro is o Kareficuvov 

 o-vXXvnevpevoi tu> Trddei, GildersleeveH says there is overlapping action. But 



* Cf. Veitch, Greek Verbs, 1887, s. v. etpo/nai : " aorist late, if correct." Kallen- 

 berg would, I suppose, regard this as an aorist. See above, p. 190. 



t I have found the imperfect in Herodotus but once (4, 125), though I have not 

 made an exhaustive search. 



| Am. Jour. Phil., 3 (1882), p. 515. § Op. cit., p. 76. 



|| Fuchs, op. cit., p. 69, cites only five of these cases. 



T Am. Jour. Phil., 24 (1903), p. 400. 



